Abstract: The topological reconstruction for the triangular faces of STL model is one of the preconditions for the surface division for some surfaces of the part model in the On-Machine Verification System（OMVS）.On the base of the topological relation between vectors and edges，the half-edge model based on the comparison of floating numbers which are the coordination of vectors in STL model are transformed into integer numbers matching based on the indexes in this paper.A new algorithm of topological reconstruction for triangular mesh based on Correlation-Hashing is proposed.In this algorithm，the hash table is used to reduce the amount of calculation for vectors merging and edges merging.By using the data correlation in STL models，the difficulty of construction hash function is reduced and the efficiency of disposing collisions is improved in the algorithm.The redundant data are filtered in the algorithm，meanwhile，the topological relations are built.The calculation examples demonstrate that the algorithm is efficient and stable.

Key words: On-Machine Verification（OMV）, topological reconstruction, triangular mesh, Stereo Lithographic（STL） model, Correlation-Hashing

摘要： 对STL（Stereo Lithographic）模型中三角面片进行拓扑重建是解决在机检测系统中对零件模型表面进行曲面划分的前提条件之一。在建立点-边拓扑关系的基础上，将原有基于顶点坐标值浮点数比较的半边匹配结构转化为基于索引值的整型数匹配，并提出了基于关联-散列结构三角网格拓扑重建方法。该方法采用散列结构为辅助数据结构来减少STL模型顶点归并和边界归并的计算量，利用STL模型相邻网格单元数据相关性降低哈希函数的构建难度并提高冲突处理效率，在滤除冗余数据的同时完成拓扑关系的建立。计算实例验证了该算法的高效性和有效性。

关键词: 在机检测, 拓扑重建, 三角网格, STL模型, 关联-散列结构